302 MR. JOHN LEIGH ON THE 



ments cannot be performed over this metal with chlorine. 

 I have prepared chlorine from perfectly pure material, in 

 vessels completly filled v^ith fluid so as to exclude all air, 

 and collected the gas over boiling water; and still, on acting 

 upon it by an absorptive solution, there was always a re- 

 siduum left. I collected the residuum from several opera- 

 tions, and analysed it. I found it to be composed entirely 

 of atmospheric air. Unless the exact proportion of this be 

 accurately determined, it will vitiate the results, as in Dr. 

 Fyfe's experiments, because it must be allowed for, which 

 has not generally been done. But, unfortunately, the pro- 

 portion of this unabsorbable residue varies with every por- 

 tion collected; and, therefore, the examination of one bottle 

 of chlorine does not give the true amount of impurity to be 

 allowed for in that employed for the analysis. From the 

 same quantity of gas (chlorine) I obtained in one instance, 

 •625; in another, '637; in another, "03, as the amount of 

 residuum. I was long puzzled as to the source of this 

 impurity; but at length found it to proceed from the air, 

 mechanically retained or absorbed by the water and other 

 fluids employed in the processes. This was displaced by 

 the more absorbable chlorine, and hence its constant pre- 

 sence in the unabsorbed chlorine, and the constant varia- 

 tion in its quality. Of late, Professor Bunsen of Marburg, 

 finding the difficulty of truly determining the amount of 

 olefiant gas by the means hitherto employed, has proposed 

 and adopted the use of charcoal or coke balls, saturated with 

 fuming sulphuric acid, for the removal of the olefiant gas. 

 This process fully answered for the separation of the illu- 

 minating gases; but the experiments, and the whole of the 

 subsequent analysis of the residual gas, have to be performed 

 over mercury, and require great care in the manipulations; 

 still it is by much the best hitherto proposed. The best 

 method of determining the general value of the gas is also 

 due to the late Dr. Henry, and is the one which I have 



